However, it does not automatically follow that Gould was "clear, consistent, and unambiguous" on the subject, as Jason Rosenhouse (and others) have suggested. At least his thinking was never that clear to me. I hate to admit it, but I think that Gould was indeed something of an "equivocator" as Dembski suggests (without noticing the irony of such a charge, apparently). Since others seem to be oblivious to this aspect of Gould's thinking, I should illustrate what I have in mind.
Here is Gould at the height of his powers, writing about saltation in his classic "Is a new and general theory of evolution emerging?" (Paleobiology 6, 119-30; 1980):
[...] the other alternative, treated with caution, reluctance, disdain or even fear by the modern synthesis, now deserves a rehearing in the light of renewed interest in development: perhaps, in many cases, the intermediates never existed. I do not refer to the saltational origin of entire new designs, complete in all their complex and integrated features -- a fantasy that would be truly anti-Darwinian in denying any creativity to selection and relegating it to the role of eliminating old models. Instead, I envisage a potential saltational origin for the essential features of key adaptations. [...] Yet Darwin, conflating gradualism with natural selection as he did so often, wrongly proclaimed that any such discontinuity, even for organs (much less taxa) would destroy his theory.(Note how he directly contradicts the statement attributed to him by Pivar.) This is typical of Gould's writings of the time, where he tended to muddle the time scale over which to apply his punctuationism and Darwin's natura non facit saltum. In the end, we're left wondering over exactly what is being proposed that would require the emergence of a "new and general theory", if it isn't Goldschmidtian "hopeful monsterism" (I sense a new religion coming...).
In the same paper he also discussed whether macroevolution can be explained by extrapolating from microevolutionary processes, another of Gould's leitmotivs from that period:
But if species originate in geological instants and then do not alter in major ways, then evolutionary trends cannot represent a simple extrapolation of allelic substitution within a population. Trends must be the product of differential success among species. In other words, species themselves must be inputs, and trends the result of their differential origin and survival. Speciation interposes itself as an irreducible level between change in local populations and trends in geological time. Macroevolution is, as Stanley argues, decoupled from microevolution. [...](My emphasis. Bibliographic references removed.)
As a final point about the extrapolation of methods for the study of events within populations, the cladogenetic basis of macroevolution virtually precludes any direct application of the primary apparatus for microevolutionary theory: classical population genetics. I believe that essentially all macroevolution is cladogenesis and its concatenated effects. What we call 'anagenesis' and often attempt to delineate as a separate phyletic process leading to 'progress,' is just accumulated cladogenesis filtered through the directing force of species selection [...]
Not exactly an endorsement of natural selection's ability to shape macroevolution, is it?